Treatment of Extended Kalman Filter Implementations for the Gyroless Star Tracker

Area of Science:

• Aerospace Engineering
• Astrodynamics
• Estimation Theory

Background:

• Traditional attitude filtering often involves multiple sensors (inertial and state-measuring devices).
• Recent advancements in star trackers enable sole-device attitude determination, leveraging unbiased stellar measurements and improved optical sensor performance.
• Estimation theory classifies filters into additive and multiplicative approaches based on quaternion update methods.

Purpose of the Study:

• To implement and compare additive and multiplicative attitude filtering techniques for a single star tracker system.
• To evaluate the performance of these filters using both simulated and real night sky image data.
• To establish a baseline for comparison using a linear least squares estimate.

Main Methods:

• Implementation of additive and multiplicative attitude filtering algorithms for a sole star tracker.
• Utilizing simulated and real night sky image datasets for filter evaluation.
• Comparison of filter results against a linear least squares estimation baseline.

Main Results:

• The state-based approach demonstrated superior accuracy and precision in estimating angular velocity compared to error-based filters.
• No significant difference was observed between the additive and multiplicative techniques for attitude determination (pointing accuracy).
• Both simulated and real data analyses yielded consistent findings regarding filter performance.

Conclusions:

• The state approach offers enhanced angular velocity estimation for single-device attitude systems.
• Attitude determination accuracy is comparable between additive and multiplicative filters for sole star tracker applications.
• Accurate attitude and rate estimation are critical for space situational awareness and object localization.